1 /* notmuch - Not much of an email program, (just index and search)
3 * Copyright © 2009 Carl Worth
5 * This program is free software: you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation, either version 3 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License
16 * along with this program. If not, see http://www.gnu.org/licenses/ .
18 * Author: Carl Worth <cworth@cworth.org>
21 #include "notmuch-client.h"
25 typedef struct _filename_node {
27 struct _filename_node *next;
30 typedef struct _filename_list {
31 _filename_node_t *head;
32 _filename_node_t **tail;
38 const char **new_tags;
39 size_t new_tags_length;
44 struct timeval tv_start;
46 _filename_list_t *removed_files;
47 _filename_list_t *removed_directories;
50 static volatile sig_atomic_t do_add_files_print_progress = 0;
53 handle_sigalrm (unused (int signal))
55 do_add_files_print_progress = 1;
58 static volatile sig_atomic_t interrupted;
61 handle_sigint (unused (int sig))
64 static char msg[] = "Stopping... \n";
66 ignored = write(2, msg, sizeof(msg)-1);
70 static _filename_list_t *
71 _filename_list_create (const void *ctx)
73 _filename_list_t *list;
75 list = talloc (ctx, _filename_list_t);
80 list->tail = &list->head;
86 _filename_list_add (_filename_list_t *list,
89 _filename_node_t *node = talloc (list, _filename_node_t);
91 node->filename = talloc_strdup (list, filename);
95 list->tail = &node->next;
99 add_files_print_progress (add_files_state_t *state)
101 struct timeval tv_now;
102 double elapsed_overall, rate_overall;
104 gettimeofday (&tv_now, NULL);
106 elapsed_overall = notmuch_time_elapsed (state->tv_start, tv_now);
107 rate_overall = (state->processed_files) / elapsed_overall;
109 printf ("Processed %d", state->processed_files);
111 if (state->total_files) {
112 double time_remaining;
114 time_remaining = ((state->total_files - state->processed_files) /
116 printf (" of %d files (", state->total_files);
117 notmuch_time_print_formatted_seconds (time_remaining);
118 printf (" remaining). \r");
120 printf (" files (%d files/sec.) \r", (int) rate_overall);
127 dirent_sort_inode (const struct dirent **a, const struct dirent **b)
129 return ((*a)->d_ino < (*b)->d_ino) ? -1 : 1;
133 dirent_sort_strcmp_name (const struct dirent **a, const struct dirent **b)
135 return strcmp ((*a)->d_name, (*b)->d_name);
138 /* Test if the directory looks like a Maildir directory.
140 * Search through the array of directory entries to see if we can find all
141 * three subdirectories typical for Maildir, that is "new", "cur", and "tmp".
143 * Return 1 if the directory looks like a Maildir and 0 otherwise.
146 _entries_resemble_maildir (struct dirent **entries, int count)
150 for (i = 0; i < count; i++) {
151 if (entries[i]->d_type != DT_DIR && entries[i]->d_type != DT_UNKNOWN)
154 if (strcmp(entries[i]->d_name, "new") == 0 ||
155 strcmp(entries[i]->d_name, "cur") == 0 ||
156 strcmp(entries[i]->d_name, "tmp") == 0)
167 /* Examine 'path' recursively as follows:
169 * o Ask the filesystem for the mtime of 'path' (fs_mtime)
170 * o Ask the database for its timestamp of 'path' (db_mtime)
172 * o Ask the filesystem for files and directories within 'path'
173 * (via scandir and stored in fs_entries)
174 * o Ask the database for files and directories within 'path'
175 * (db_files and db_subdirs)
177 * o Pass 1: For each directory in fs_entries, recursively call into
178 * this same function.
180 * o Pass 2: If 'fs_mtime' > 'db_mtime', then walk fs_entries
181 * simultaneously with db_files and db_subdirs. Look for one of
182 * three interesting cases:
184 * 1. Regular file in fs_entries and not in db_files
185 * This is a new file to add_message into the database.
187 * 2. Filename in db_files not in fs_entries.
188 * This is a file that has been removed from the mail store.
190 * 3. Directory in db_subdirs not in fs_entries
191 * This is a directory that has been removed from the mail store.
193 * Note that the addition of a directory is not interesting here,
194 * since that will have been taken care of in pass 1. Also, we
195 * don't immediately act on file/directory removal since we must
196 * ensure that in the case of a rename that the new filename is
197 * added before the old filename is removed, (so that no
198 * information is lost from the database).
200 * o Tell the database to update its time of 'path' to 'fs_mtime'
202 static notmuch_status_t
203 add_files_recursive (notmuch_database_t *notmuch,
205 add_files_state_t *state)
208 struct dirent *entry = NULL;
210 time_t fs_mtime, db_mtime;
211 notmuch_status_t status, ret = NOTMUCH_STATUS_SUCCESS;
212 notmuch_message_t *message = NULL;
213 struct dirent **fs_entries = NULL;
214 int i, num_fs_entries;
215 notmuch_directory_t *directory;
216 notmuch_filenames_t *db_files = NULL;
217 notmuch_filenames_t *db_subdirs = NULL;
219 notmuch_bool_t is_maildir, new_directory;
222 if (stat (path, &st)) {
223 fprintf (stderr, "Error reading directory %s: %s\n",
224 path, strerror (errno));
225 return NOTMUCH_STATUS_FILE_ERROR;
228 /* This is not an error since we may have recursed based on a
229 * symlink to a regular file, not a directory, and we don't know
230 * that until this stat. */
231 if (! S_ISDIR (st.st_mode))
232 return NOTMUCH_STATUS_SUCCESS;
234 fs_mtime = st.st_mtime;
236 directory = notmuch_database_get_directory (notmuch, path);
237 db_mtime = notmuch_directory_get_mtime (directory);
240 new_directory = TRUE;
244 new_directory = FALSE;
245 db_files = notmuch_directory_get_child_files (directory);
246 db_subdirs = notmuch_directory_get_child_directories (directory);
249 /* If the database knows about this directory, then we sort based
250 * on strcmp to match the database sorting. Otherwise, we can do
251 * inode-based sorting for faster filesystem operation. */
252 num_fs_entries = scandir (path, &fs_entries, 0,
254 dirent_sort_inode : dirent_sort_strcmp_name);
256 if (num_fs_entries == -1) {
257 fprintf (stderr, "Error opening directory %s: %s\n",
258 path, strerror (errno));
259 ret = NOTMUCH_STATUS_FILE_ERROR;
263 /* Pass 1: Recurse into all sub-directories. */
264 is_maildir = _entries_resemble_maildir (fs_entries, num_fs_entries);
266 for (i = 0; i < num_fs_entries; i++) {
270 entry = fs_entries[i];
272 /* We only want to descend into directories.
273 * But symlinks can be to directories too, of course.
275 * And if the filesystem doesn't tell us the file type in the
276 * scandir results, then it might be a directory (and if not,
277 * then we'll stat and return immediately in the next level of
279 if (entry->d_type != DT_DIR &&
280 entry->d_type != DT_LNK &&
281 entry->d_type != DT_UNKNOWN)
286 /* Ignore special directories to avoid infinite recursion.
287 * Also ignore the .notmuch directory and any "tmp" directory
288 * that appears within a maildir.
290 /* XXX: Eventually we'll want more sophistication to let the
291 * user specify files to be ignored. */
292 if (strcmp (entry->d_name, ".") == 0 ||
293 strcmp (entry->d_name, "..") == 0 ||
294 (is_maildir && strcmp (entry->d_name, "tmp") == 0) ||
295 strcmp (entry->d_name, ".notmuch") ==0)
300 next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
301 status = add_files_recursive (notmuch, next, state);
302 if (status && ret == NOTMUCH_STATUS_SUCCESS)
308 /* If this directory hasn't been modified since the last
309 * "notmuch new", then we can skip the second pass entirely. */
310 if (fs_mtime <= db_mtime)
313 /* Pass 2: Scan for new files, removed files, and removed directories. */
314 for (i = 0; i < num_fs_entries; i++)
319 entry = fs_entries[i];
321 /* Check if we've walked past any names in db_files or
322 * db_subdirs. If so, these have been deleted. */
323 while (notmuch_filenames_valid (db_files) &&
324 strcmp (notmuch_filenames_get (db_files), entry->d_name) < 0)
326 char *absolute = talloc_asprintf (state->removed_files,
328 notmuch_filenames_get (db_files));
330 _filename_list_add (state->removed_files, absolute);
332 notmuch_filenames_move_to_next (db_files);
335 while (notmuch_filenames_valid (db_subdirs) &&
336 strcmp (notmuch_filenames_get (db_subdirs), entry->d_name) <= 0)
338 const char *filename = notmuch_filenames_get (db_subdirs);
340 if (strcmp (filename, entry->d_name) < 0)
342 char *absolute = talloc_asprintf (state->removed_directories,
343 "%s/%s", path, filename);
345 _filename_list_add (state->removed_directories, absolute);
348 notmuch_filenames_move_to_next (db_subdirs);
351 /* If we're looking at a symlink, we only want to add it if it
352 * links to a regular file, (and not to a directory, say).
354 * Similarly, if the file is of unknown type (due to filesytem
355 * limitations), then we also need to look closer.
357 * In either case, a stat does the trick.
359 if (entry->d_type == DT_LNK || entry->d_type == DT_UNKNOWN) {
362 next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
363 err = stat (next, &st);
367 /* Don't emit an error for a link pointing nowhere, since
368 * the directory-traversal pass will have already done
373 if (! S_ISREG (st.st_mode))
375 } else if (entry->d_type != DT_REG) {
379 /* Don't add a file that we've added before. */
380 if (notmuch_filenames_valid (db_files) &&
381 strcmp (notmuch_filenames_get (db_files), entry->d_name) == 0)
383 notmuch_filenames_move_to_next (db_files);
387 /* We're now looking at a regular file that doesn't yet exist
388 * in the database, so add it. */
389 next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name);
391 state->processed_files++;
393 if (state->verbose) {
394 if (state->output_is_a_tty)
398 state->processed_files,
402 putchar((state->output_is_a_tty) ? '\r' : '\n');
406 status = notmuch_database_add_message (notmuch, next, &message);
409 case NOTMUCH_STATUS_SUCCESS:
410 state->added_messages++;
411 for (tag=state->new_tags; *tag != NULL; tag++)
412 notmuch_message_add_tag (message, *tag);
414 /* Non-fatal issues (go on to next file) */
415 case NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID:
416 /* Stay silent on this one. */
418 case NOTMUCH_STATUS_FILE_NOT_EMAIL:
419 fprintf (stderr, "Note: Ignoring non-mail file: %s\n",
422 /* Fatal issues. Don't process anymore. */
423 case NOTMUCH_STATUS_READ_ONLY_DATABASE:
424 case NOTMUCH_STATUS_XAPIAN_EXCEPTION:
425 case NOTMUCH_STATUS_OUT_OF_MEMORY:
426 fprintf (stderr, "Error: %s. Halting processing.\n",
427 notmuch_status_to_string (status));
431 case NOTMUCH_STATUS_FILE_ERROR:
432 case NOTMUCH_STATUS_NULL_POINTER:
433 case NOTMUCH_STATUS_TAG_TOO_LONG:
434 case NOTMUCH_STATUS_UNBALANCED_FREEZE_THAW:
435 case NOTMUCH_STATUS_LAST_STATUS:
436 INTERNAL_ERROR ("add_message returned unexpected value: %d", status);
441 notmuch_message_destroy (message);
445 if (do_add_files_print_progress) {
446 do_add_files_print_progress = 0;
447 add_files_print_progress (state);
457 /* Now that we've walked the whole filesystem list, anything left
458 * over in the database lists has been deleted. */
459 while (notmuch_filenames_valid (db_files))
461 char *absolute = talloc_asprintf (state->removed_files,
463 notmuch_filenames_get (db_files));
465 _filename_list_add (state->removed_files, absolute);
467 notmuch_filenames_move_to_next (db_files);
470 while (notmuch_filenames_valid (db_subdirs))
472 char *absolute = talloc_asprintf (state->removed_directories,
474 notmuch_filenames_get (db_subdirs));
476 _filename_list_add (state->removed_directories, absolute);
478 notmuch_filenames_move_to_next (db_subdirs);
482 status = notmuch_directory_set_mtime (directory, fs_mtime);
483 if (status && ret == NOTMUCH_STATUS_SUCCESS)
497 notmuch_filenames_destroy (db_subdirs);
499 notmuch_filenames_destroy (db_files);
501 notmuch_directory_destroy (directory);
506 /* This is the top-level entry point for add_files. It does a couple
507 * of error checks, sets up the progress-printing timer and then calls
508 * into the recursive function. */
509 static notmuch_status_t
510 add_files (notmuch_database_t *notmuch,
512 add_files_state_t *state)
514 notmuch_status_t status;
515 struct sigaction action;
516 struct itimerval timerval;
517 notmuch_bool_t timer_is_active = FALSE;
520 if (state->output_is_a_tty && ! debugger_is_active () && ! state->verbose) {
521 /* Setup our handler for SIGALRM */
522 memset (&action, 0, sizeof (struct sigaction));
523 action.sa_handler = handle_sigalrm;
524 sigemptyset (&action.sa_mask);
525 action.sa_flags = SA_RESTART;
526 sigaction (SIGALRM, &action, NULL);
528 /* Then start a timer to send SIGALRM once per second. */
529 timerval.it_interval.tv_sec = 1;
530 timerval.it_interval.tv_usec = 0;
531 timerval.it_value.tv_sec = 1;
532 timerval.it_value.tv_usec = 0;
533 setitimer (ITIMER_REAL, &timerval, NULL);
535 timer_is_active = TRUE;
538 if (stat (path, &st)) {
539 fprintf (stderr, "Error reading directory %s: %s\n",
540 path, strerror (errno));
541 return NOTMUCH_STATUS_FILE_ERROR;
544 if (! S_ISDIR (st.st_mode)) {
545 fprintf (stderr, "Error: %s is not a directory.\n", path);
546 return NOTMUCH_STATUS_FILE_ERROR;
549 status = add_files_recursive (notmuch, path, state);
551 if (timer_is_active) {
552 /* Now stop the timer. */
553 timerval.it_interval.tv_sec = 0;
554 timerval.it_interval.tv_usec = 0;
555 timerval.it_value.tv_sec = 0;
556 timerval.it_value.tv_usec = 0;
557 setitimer (ITIMER_REAL, &timerval, NULL);
559 /* And disable the signal handler. */
560 action.sa_handler = SIG_IGN;
561 sigaction (SIGALRM, &action, NULL);
567 /* XXX: This should be merged with the add_files function since it
568 * shares a lot of logic with it. */
569 /* Recursively count all regular files in path and all sub-directories
570 * of path. The result is added to *count (which should be
571 * initialized to zero by the top-level caller before calling
574 count_files (const char *path, int *count)
576 struct dirent *entry = NULL;
579 struct dirent **fs_entries = NULL;
580 int num_fs_entries = scandir (path, &fs_entries, 0, dirent_sort_inode);
583 if (num_fs_entries == -1) {
584 fprintf (stderr, "Warning: failed to open directory %s: %s\n",
585 path, strerror (errno));
589 while (!interrupted) {
590 if (i == num_fs_entries)
593 entry = fs_entries[i++];
595 /* Ignore special directories to avoid infinite recursion.
596 * Also ignore the .notmuch directory.
598 /* XXX: Eventually we'll want more sophistication to let the
599 * user specify files to be ignored. */
600 if (strcmp (entry->d_name, ".") == 0 ||
601 strcmp (entry->d_name, "..") == 0 ||
602 strcmp (entry->d_name, ".notmuch") == 0)
607 if (asprintf (&next, "%s/%s", path, entry->d_name) == -1) {
609 fprintf (stderr, "Error descending from %s to %s: Out of memory\n",
610 path, entry->d_name);
616 if (S_ISREG (st.st_mode)) {
618 if (*count % 1000 == 0) {
619 printf ("Found %d files so far.\r", *count);
622 } else if (S_ISDIR (st.st_mode)) {
623 count_files (next, count);
637 upgrade_print_progress (void *closure,
640 add_files_state_t *state = closure;
642 printf ("Upgrading database: %.2f%% complete", progress * 100.0);
645 struct timeval tv_now;
646 double elapsed, time_remaining;
648 gettimeofday (&tv_now, NULL);
650 elapsed = notmuch_time_elapsed (state->tv_start, tv_now);
651 time_remaining = (elapsed / progress) * (1.0 - progress);
653 notmuch_time_print_formatted_seconds (time_remaining);
654 printf (" remaining)");
662 /* Recursively remove all filenames from the database referring to
663 * 'path' (or to any of its children). */
665 _remove_directory (void *ctx,
666 notmuch_database_t *notmuch,
671 notmuch_directory_t *directory;
672 notmuch_filenames_t *files, *subdirs;
673 notmuch_status_t status;
676 directory = notmuch_database_get_directory (notmuch, path);
678 for (files = notmuch_directory_get_child_files (directory);
679 notmuch_filenames_valid (files);
680 notmuch_filenames_move_to_next (files))
682 absolute = talloc_asprintf (ctx, "%s/%s", path,
683 notmuch_filenames_get (files));
684 status = notmuch_database_remove_message (notmuch, absolute);
685 if (status == NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID)
686 *renamed_files = *renamed_files + 1;
688 *removed_files = *removed_files + 1;
689 talloc_free (absolute);
692 for (subdirs = notmuch_directory_get_child_directories (directory);
693 notmuch_filenames_valid (subdirs);
694 notmuch_filenames_move_to_next (subdirs))
696 absolute = talloc_asprintf (ctx, "%s/%s", path,
697 notmuch_filenames_get (subdirs));
698 _remove_directory (ctx, notmuch, absolute, renamed_files, removed_files);
699 talloc_free (absolute);
702 notmuch_directory_destroy (directory);
706 notmuch_new_command (void *ctx, int argc, char *argv[])
708 notmuch_config_t *config;
709 notmuch_database_t *notmuch;
710 add_files_state_t add_files_state;
712 struct timeval tv_now;
716 char *dot_notmuch_path;
717 struct sigaction action;
719 int renamed_files, removed_files;
720 notmuch_status_t status;
723 add_files_state.verbose = 0;
724 add_files_state.output_is_a_tty = isatty (fileno (stdout));
726 for (i = 0; i < argc && argv[i][0] == '-'; i++) {
727 if (STRNCMP_LITERAL (argv[i], "--verbose") == 0) {
728 add_files_state.verbose = 1;
730 fprintf (stderr, "Unrecognized option: %s\n", argv[i]);
735 config = notmuch_config_open (ctx, NULL, NULL);
739 add_files_state.new_tags = notmuch_config_get_new_tags (config, &add_files_state.new_tags_length);
740 db_path = notmuch_config_get_database_path (config);
742 dot_notmuch_path = talloc_asprintf (ctx, "%s/%s", db_path, ".notmuch");
744 if (stat (dot_notmuch_path, &st)) {
748 count_files (db_path, &count);
752 printf ("Found %d total files (that's not much mail).\n", count);
753 notmuch = notmuch_database_create (db_path);
754 add_files_state.total_files = count;
756 notmuch = notmuch_database_open (db_path,
757 NOTMUCH_DATABASE_MODE_READ_WRITE);
761 if (notmuch_database_needs_upgrade (notmuch)) {
762 printf ("Welcome to a new version of notmuch! Your database will now be upgraded.\n");
763 gettimeofday (&add_files_state.tv_start, NULL);
764 notmuch_database_upgrade (notmuch, upgrade_print_progress,
766 printf ("Your notmuch database has now been upgraded to database format version %u.\n",
767 notmuch_database_get_version (notmuch));
770 add_files_state.total_files = 0;
776 /* Setup our handler for SIGINT. We do this after having
777 * potentially done a database upgrade we this interrupt handler
779 memset (&action, 0, sizeof (struct sigaction));
780 action.sa_handler = handle_sigint;
781 sigemptyset (&action.sa_mask);
782 action.sa_flags = SA_RESTART;
783 sigaction (SIGINT, &action, NULL);
785 talloc_free (dot_notmuch_path);
786 dot_notmuch_path = NULL;
788 add_files_state.processed_files = 0;
789 add_files_state.added_messages = 0;
790 gettimeofday (&add_files_state.tv_start, NULL);
792 add_files_state.removed_files = _filename_list_create (ctx);
793 add_files_state.removed_directories = _filename_list_create (ctx);
795 ret = add_files (notmuch, db_path, &add_files_state);
799 for (f = add_files_state.removed_files->head; f; f = f->next) {
800 status = notmuch_database_remove_message (notmuch, f->filename);
801 if (status == NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID)
807 for (f = add_files_state.removed_directories->head; f; f = f->next) {
808 _remove_directory (ctx, notmuch, f->filename,
809 &renamed_files, &removed_files);
812 talloc_free (add_files_state.removed_files);
813 talloc_free (add_files_state.removed_directories);
815 gettimeofday (&tv_now, NULL);
816 elapsed = notmuch_time_elapsed (add_files_state.tv_start,
819 if (add_files_state.processed_files) {
820 printf ("Processed %d %s in ", add_files_state.processed_files,
821 add_files_state.processed_files == 1 ?
822 "file" : "total files");
823 notmuch_time_print_formatted_seconds (elapsed);
825 printf (" (%d files/sec.). \n",
826 (int) (add_files_state.processed_files / elapsed));
832 if (add_files_state.added_messages) {
833 printf ("Added %d new %s to the database.",
834 add_files_state.added_messages,
835 add_files_state.added_messages == 1 ?
836 "message" : "messages");
838 printf ("No new mail.");
842 printf (" Removed %d %s.",
844 removed_files == 1 ? "message" : "messages");
848 printf (" Detected %d file %s.",
850 renamed_files == 1 ? "rename" : "renames");
856 printf ("\nNote: At least one error was encountered: %s\n",
857 notmuch_status_to_string (ret));
860 notmuch_database_close (notmuch);
862 return ret || interrupted;